Alternative to HRAs: Using microsegmentation to reach at-risk patients
Tobacco and food companies find potential customers by using demographic and consumer attitude information, but those kinds of data are usually not used by healthcare companies.
Instead, most health insurers and population health companies try to find at-risk members through health risk assessments (HRA) and medical claims information.
HRAs ask members and employees about their health status and their desire to change their lifestyle habits. Although HRAs find individuals who are ready to improve their health, many people don’t fill out HRAs because they feel the information is intrusive and that their employer or health plan will use it against them. This often leads to low participation rates.
As a way to more effectively find at-risk people and increase engagement rates in wellness programs, Boston-based population health company Health Dialog created the Wellness SegmenterTM. The challenge is that health and wellness outreach efforts are usually either broad-based and poorly targeted or focus only on the few people who fill out an HRA, making it difficult for health plans and employers to engage individuals who may benefit from health and wellness programs, says Elizabeth Barbeau, ScD, MPH, senior vice president at Health Dialog.
Many employers, population health companies, and health plans use HRAs to find people who need intervention, but Barbeau says the assessments often don’t find at-risk people. In fact, 40% of those who are identified as smokers by their providers on a medical claim do not self-report smoking on HRAs. And that doesn’t even include the people who refuse to fill out an HRA.
Although adding incentives for completing HRAs increases compliance, Barbeau says that doesn’t mean that individuals are truthful on them, particularly if the HRA responses, such as smoking status, are used to determine qualifications for benefits, such as reduced premiums.
HRAs provide important information but are insufficient to target a total population, Barbeau says.
“We’re spending way too much money on HRAs to solely identify individuals,” she says. “There are other ways to do this, and we can look outside our industry to do it.”
The issue is threefold: few people bother to fill out an HRA, those who do aren’t always truthful, and medical claims often don’t properly identify obesity and smokers. This means the net is too small and not effective.
Health Dialog’s solution uses predictive models that leverage third-party demographic and consumer data to more effectively identify and target members for health and wellness outreach interventions, says Barbeau.
The idea is to move away from the so-called “spray and pray” method, which means reaching out to all members regardless of risk profile. This means you provide a broad audience with generic messaging, which is expensive and ineffective. And this method sidesteps the “cherry-picking” that comes with only targeting the small percentage of people who fill out an HRA, says Barbeau.
Instead, Health Dialog promotes the idea of using third-party demographic and consumer data to find people who are most likely at risk of specific diseases or lifestyle issues, such as smokers; reaching out to that group; and personalizing the message. Barbeau studied the way the tobacco industry reaches its client base and found some parallels for healthcare companies looking to engage patients.
The underlying premise is that there are specific patterns of demographic characteristics, consumer behaviors, and lifestyle-related medical conditions, which, when taken together, can predict lifestyle behaviors and risks, says Barbeau.
The Wellness Segmenter is a suite of lifestyle-based predictive models that identify individuals with a high probability of being a current smoker or obese. Although obesity is a national problem, obesity rates are higher in some regions. Those are the areas in which the Wellness Segmenter could find possible at-risk members.
In this model, Health Dialog uses community data, such as age, consumer spending, language/ethnicity, and education level; claims data, such as costs and procedures; utilization, such as provider visits and types, ER visits, and frequency of visits; and additional sources, such as HRA and personal health record data.
Barbeau says Health Dialog microsegments populations by using ZIP code plus four, which corresponds to approximately 16 households, and combines it with the other information specific to these microsegments. Each individual is assigned a risk score for obesity and smoking.
“What is really important in this is that it can all be done in the absence of the HRA data,” explains Barbeau. “We can then identify a stratified list of individuals who are likely to fall into those categories [smokers and obese].”
When Health Dialog built the models, the company let the empirical data tell it how to weigh claims, community, utilization, and additional sources to come up with the best prediction. “We created a laundry list from medical claims and consumer segmentation data that we thought would likely be associated with an individual identifying on an HRA and rigorously evaluated which ones have the highest predictive value in identifying obesity and smoking,” says Barbeau.
Health Dialog has found that by using the lifestyle-based predictive model for obesity, the company doubles its chances of correctly identifying obese individuals. By zeroing in on a key 15% of the population, Health Dialog was able to more accurately find at-risk patients than by simply using HRAs, according to the company. (See Figure 10 below and Figure 11 on p. 12.)
“It’s a pretty powerful indicator,” says Barbeau about the Wellness Segmenter.
With that information, the company is also able to target resources to reach out to members with a personalized and meaningful message that resonates with the individuals’ lifestyles and will lead to higher engagements, she says.
Barbeau adds that the Wellness Segmenter could actually help reduce racial disparities in healthcare.
“I think one of the things that got me so excited when I joined Health Dialog about this consumer model is it represents an effective way to tackle health disparities,” she says. “If we keep doing the same health behavior and health interventions, we are going to see an exacerbation of health disparities by race, income, and educational attainment. We need to find ways to identify and engage all members of a population, including those who public and commercial entities have had a hard time reaching to date.”
Health Net tries Wellness Segmenter
Health Net, the Woodland, CA–based health insurer with more than 6 million members, is an example of a health plan eager to sharpen its wellness program participation rates. Although Health Net has traditionally used HRAs and set up numerous avenues to reach out to members, it is stepping up efforts to bolster participation rates.
Jennifer Christian-Herman, PhD, vice president of national integrated health improvement at Health Net, says that in the past, the insurer has tried to find at-risk members through HRAs, claims data, and laboratory data, but it still can be difficult to pinpoint the right individuals.
“One of the things we found was that by doing broad outreach without targeting the specific population, you are not necessarily identifying the people who would benefit most from a particular wellness program,” says Christian-Herman.
One of Health Net’s key areas of interest is supporting individuals with cardiometabolic risk, which is a combination of risk factors (including smoking; blood pressure; key lab values, such as fasting blood sugar, triglycerides, and HDL cholesterol; and girth measurement) that strongly predict future heart disease, diabetes, and other preventable health conditions. The insurer has worked with Health Dialog, and by now integrating the Wellness Segmenter, the insurer and population health company broadened their relationship.
Christian-Herman says Health Net is interested in identifying its patient population with a cardiometabolic risk and performing targeted outreach. In the short-term, Health Net will gauge smoking cessation rates and weight loss and track lab results. In the long-term, the insurer will track whether the program is preventing the clinical outcomes associated with metabolic syndrome.
Christian-Herman says the program is part of a larger integrated whole-person approach. “An important aspect of our Decision Power Health and Wellness program is offering multiple ways to reach out to people,” she says. “We have integrated our Web strategy and our online tools with our health coaching and telephonic programs. We know that people have multiple ways of learning and choosing to engage in behavior change, and we try to meet them where they are clinically and support them in how they would like to learn or to change.”
- CMS Seeks to 'Rapidly Reduce' Medicare Spending with $1B in Grants
- Building a Better Healthcare Board
- Case Study: Advance Care Conversations
- Patient Harm Data to Remain on Medicare's Hospital Compare Site
- Quiet ORs Better for Patient Safety
- Hard-Nosed About Physician Teamwork
- Hospital Pricing Data Dump Won't Hurt You, Yet
- CMS Releases Hospital Pricing Data
- Tavenner Confirmed as CMS Administrator
- Evidence-Based Practice and Nursing Research: Avoiding Confusion